Emerging Technology
NVIDIA's GTC 2025: Top Reveals, Industry Reaction, & More
By Manali Kekade
Updated on Wed, Mar 19, 2025
Share
Considering NVIDIA’s CEO Jensen Huang’s keynote speech at CES 2025 in January was sold out, it’s no surprise that his keynote speech at NVIDIA’s GTC was sold out too.
NVIDIA GTC is the company’s annual developers conference. GTC stands for GPU Technology Conference, where the full form of GPU is graphics processing unit. GPUs are the largest selling commodity in the artificial intelligence (AI) world right now, as they power the training and development of AI models.
The 2025 edition of NVIDIA’s GTC is being held at the San Jose McEnery Convention Center (San Jose, California), from March 17 to 21, 2025, and can be viewed virtually. While the keynote speech was scheduled for March 18, the event’s exhibits are on show from March 18 to 21, with workshops running from March 16-20.
NVIDIA also developed a Mobile App (iOS and Android) for “for the Ultimate GTC Experience.”
The event was sponsored by the likes of AWS, Dell, Google, Hewlett Packard Enterprises, Microsoft, and Azure—who made the “Elite” group—and Asus, Accenture, Asus, EY, Lenovo, and others—from the “Diamond” sponsors list.
Through the event—and Jensen Huang’s keynote speech—the company reveals what’s next in the world of AI, along with a bunch of new technologies, products, partnerships, enhancements, and more, across a wide range of industries. GTC is also a popular space for other companies to showcase their products that are integrated with NVIDIA technologies.
Founder and CEO Huang began the keynote by introducing the audience to NVIDIA’s headquarters through a stunning visual of its lobby, following which he spoke about NVIDIA’s 25-year history, its growth, and its present position.
A popular word uttered 53 times during the speech was “token”—which powers how every AI application’s algorithm works. Tokens are tiny units of data that come from breaking down bigger chunks of information, and to help people better understand this, NVIDIA published a blog post a day before the speech, explaining all one needs to know about tokens and tokenization.
He then spoke about the emergence of Agentic AI, AI’s ability to reason “step by step by step,” and how inference and reinforcement learning are driving demand for AI computing. With the demand for GPUs surging, he said AI computing is going through an “inflection point,” resulting in the value of data center buildout reaching $1 trillion. He believes every company will have two factories: one for what they build and another for AI.
Since it was introduced, the world has been in awe of NVIDIA’s Blackwell series of GPUs. So, of course, Jensen Huang couldn’t pass up the opportunity to talk about it. The good news is that the NVIDIA Blackwell platform is in full production, leading to Huang exclaiming, “How is this not beautiful?”
NVIDIA also revealed its next GPU—Blackwell Ultra—which is aimed at paving the way for the age of AI reasoning. It’s expected to be available in the second half of the year.
This also includes the introduction of the NVIDIA GB300 NVL72, which connects 72 Blackwell Ultra GPUs and 36 Arm Neoverse-based NVIDIA Grace CPUs in a rack-scale design, forming a single massive GPU. It outperforms the NVIDIA GB200 NVL72, delivering 1.5x more AI performance, and increasing revenue opportunity by 50x for AI factories.
“We designed Blackwell Ultra for this moment—it’s a single versatile platform that can easily and efficiently do pretraining, post-training, and reasoning AI inference,” said Huang.
The company also unveiled the NVIDIA DGX SuperPOD, which is built with NVIDIA Blackwell Ultra GPUs and makes the world’s most advanced enterprise AI infrastructure. It brings enterprises AI factory supercomputing with superior agentic AI reasoning. Huang said, “The NVIDIA Blackwell Ultra DGX SuperPOD provides out-of-the-box AI supercomputing for the age of agentic and physical AI.”
Paying tribute to astronomer Vera Rubin, Huang spoke about the upcoming next-generation NVIDIA Rubin Ultra GPU and NVIDIA Vera CPU architectures, which when paired together to form the Vera Rubin computing system, are expected to outperform the Blackwell architecture. Vera Rubin will come in the second half of 2026, with Vera Rubin Ultra coming in 2027.
Going one step further, Huang said the Vera Rubin system will be followed by the Feynman architecture (2028).
Describing Blackwell’s capabilities of supporting extreme scale-up he said, “The reason why we wanted to do this is to solve an extreme problem, and it’s called inference.”
To further accelerate inference on a large scale, the company is bringing NVIDIA Dynamo, an open-source modular inference framework to help accelerate and scale AI reasoning models in AI factories. It comes with the lowest cost and the highest efficiency. Huang said, “To enable a future of custom reasoning AI, NVIDIA Dynamo helps serve these models at scale, driving cost savings and efficiencies across AI factories.”
The company announced the NVIDIA RTX PRO Blackwell series, which consists of powerful workstation and server GPUs with boosted computing, AI inference, ray tracing, and neural rendering technologies, specially designed for AI, technical, creative, engineering, and design professionals. The lineup includes data center, desktop, and laptop GPUs.
Its features include NVIDIA streaming multiprocessor, fourth-generation RT cores, fifth-generation Tensor cores, larger and faster GDDR7 memory, Multi-Instance GPU (MIG), and more. The GPUs are expected to be available soon through various distribution partners, including Cisco, Dell, HP, Lenovo, and others.
Using NVIDIA NIM microservices—its optimized AI software—allow users to unlock greater possibilities and performance, allowing developers to build AI-powered app assistants, productivity tools, and content-creation workflows that integrate seamlessly.
NVIDIA’s Project DIGITS is finally here—with a new name!
The company unveiled two personal AI supercomputers—DGX Spark and DGX Station—which are powered by the NVIDIA Blackwell Ultra platform. These machines, which aim to enable AI developers, researchers, data scientists, and students to prototype, fine-tune, and inference large models on desktops, will be developed by ASUS, Dell, HP Inc., and Lenovo.
DGX Spark is the world’s smallest AI supercomputer, while DGX Station brings data center-level performance to desktops for AI development.
NVIDIA revealed it’s building a research center in Boston, Massachusetts, to provide cutting-edge technologies to advance quantum computing. It’s called the NVIDIA Accelerated Quantum Research Center, or NVAQC, and will integrate quantum hardware with AI supercomputers to solve quantum computing’s most challenging problems. “Quantum computing will augment AI supercomputers to tackle some of the world’s most important problems, from drug discovery to materials development,” said Jensen Huang.
Announcing Isaac GR00T N1, the world’s first open, customizable humanoid robot foundation model for generalized humanoid reasoning and skills, Jensen Huang proclaimed, “The age of generalist robotics is here!”
GR00T N1 forms the first of a family of fully customizable models that will be made available to robotics developers worldwide to help compensate for the shortage of skilled workers. “With NVIDIA Isaac GR00T N1 and new data-generation and robot-learning frameworks, robotics developers everywhere will open the next frontier in the age of AI.”
GR00T N1 comes with a dual-system architecture: “System 1” a fast-thinking action model to mirror human reflexes or intuition, and “System 2” a slow-thinking model for deliberate, methodical decision-making.
Adding to this is the NVIDIA Isaac GR00T Blueprint for generating synthetic data and Newton, an open-source physics engine for developing robots that is under development with Google DeepMind and Disney Research, which was exhibited on stage by Jensen Huang.
NVIDIA and GM will collaborate to build next-generation vehicles, factories and robots using AI, simulation and AI chips. This will include building custom AI systems using NVIDIA accelerated compute platforms, to optimize GM’s factory planning and robotics, as well as enhance in-vehicle hardware for future advanced driver-assistance systems (ADAS) for safe driving experiences.
This includes the NVIDIA Omniverse, a platform of APIs, SDKs, and services that enhance simulation workflows for industrial and robotic use cases. The Omniverse blueprint advances AI factory design and simulation.
“The era of physical AI is here, and together with GM, we’re transforming transportation, from vehicles to the factories where they’re made,” said Huang.
As such, NVIDIA is also releasing a massive, open-source dataset for building the next generation of physical AI, to help developers and researchers teach autonomous robots and vehicles how to interact with the physical world. Available on Hugging Face, the dataset offers developers 15 terabytes of data representing more than 320,000 trajectories for robotics training.
NVIDIA’s Halos operates at the intersection of autonomy and robotics, with a focus on the development of autonomous vehicles (AVs). It forms a comprehensive safety system combining automotive hardware and software safety solutions. From chips and software to tools and services, the holistic safety system works on three different levels.
“At the technology level, it spans platform, algorithmic, and ecosystem safety. At the development level, it includes design-time, deployment-time, and validation-time guardrails. And at the computational level, it spans AI training to deployment.” The system connects NVIDIA DGX for AI training, NVIDIA Omniverse and NVIDIA Cosmos for simulation, and NVIDIA DRIVE AGX for deployment.
It combines safety-focused technology research, development, deployment, partnerships, and collaborations by NVIDIA, including 15,000+ engineering years invested in vehicle safety, 10,000+ hours of contributions to international standards committees, 1,000+ AV-safety patents filed, and more.
“Halos’ holistic approach to safety is particularly critical in a setting where companies want to harness the power of generative AI for increasingly capable AV systems developed end to end, which preclude traditional compositional design and verification,” said Marco Pavone, lead AV researcher at NVIDIA.
Building on their previous partnership, NVIDIA, Alphabet and Google revealed new initiatives to advance AI, democratize access to AI tools, speed up the development of physical AI, and transform industries including healthcare, manufacturing, and energy. This will include using AI and simulation technology to develop robots with grasping skills, reimagine drug discovery, optimize energy grids, and more.
“I’m really excited about the next phase of our partnership as we work together on agentic AI, robotics, and bringing the benefits of AI to more people around the world,” said Sundar Pichai, CEO of Google and Alphabet.
“It’s a great joy to see Google and NVIDIA researchers and engineers collaborate to solve incredible challenges, from drug discovery to robotics,” said Jensen Huang.
Google DeepMind and NVIDIA will work together to develop responsible AI and open models.
Yum! Brands, the world’s largest restaurant chain and parent company of KFC, Taco Bell, Pizza Hut, and Habit Burger & Grill, is partnering with NVIDIA to deploy multiple AI solutions in over 500 restaurants this year. The company will also use NVIDIA computer vision software to enhance staffing solutions by analyzing drive-thru traffic and more.
Oracle and NVIDIA announced a first-of-its-kind integration between NVIDIA accelerated computing and inference software with Oracle’s AI infrastructure and generative AI services, to help organizations speed up the creation of agentic AI applications. This will be done by making 160+ AI tools and 100+ NVIDIA NIM microservices natively available through the OCI Console, among other moves.
What do you think about the reveals made so far at NVIDIA GTC 2025?
Let us know in the comments below!
What Is NVIDIA GTC?
NVIDIA GTC is the company’s annual developers conference. GTC stands for GPU Technology Conference, where the full form of GPU is graphics processing unit. GPUs are the largest selling commodity in the artificial intelligence (AI) world right now, as they power the training and development of AI models.
The 2025 edition of NVIDIA’s GTC is being held at the San Jose McEnery Convention Center (San Jose, California), from March 17 to 21, 2025, and can be viewed virtually. While the keynote speech was scheduled for March 18, the event’s exhibits are on show from March 18 to 21, with workshops running from March 16-20.
NVIDIA also developed a Mobile App (iOS and Android) for “for the Ultimate GTC Experience.”
The event was sponsored by the likes of AWS, Dell, Google, Hewlett Packard Enterprises, Microsoft, and Azure—who made the “Elite” group—and Asus, Accenture, Asus, EY, Lenovo, and others—from the “Diamond” sponsors list.
Through the event—and Jensen Huang’s keynote speech—the company reveals what’s next in the world of AI, along with a bunch of new technologies, products, partnerships, enhancements, and more, across a wide range of industries. GTC is also a popular space for other companies to showcase their products that are integrated with NVIDIA technologies.
Founder and CEO Huang began the keynote by introducing the audience to NVIDIA’s headquarters through a stunning visual of its lobby, following which he spoke about NVIDIA’s 25-year history, its growth, and its present position.
A popular word uttered 53 times during the speech was “token”—which powers how every AI application’s algorithm works. Tokens are tiny units of data that come from breaking down bigger chunks of information, and to help people better understand this, NVIDIA published a blog post a day before the speech, explaining all one needs to know about tokens and tokenization.
AI’s $1 Trillion Inflection Point
He then spoke about the emergence of Agentic AI, AI’s ability to reason “step by step by step,” and how inference and reinforcement learning are driving demand for AI computing. With the demand for GPUs surging, he said AI computing is going through an “inflection point,” resulting in the value of data center buildout reaching $1 trillion. He believes every company will have two factories: one for what they build and another for AI.
NVIDIA Blackwell
Since it was introduced, the world has been in awe of NVIDIA’s Blackwell series of GPUs. So, of course, Jensen Huang couldn’t pass up the opportunity to talk about it. The good news is that the NVIDIA Blackwell platform is in full production, leading to Huang exclaiming, “How is this not beautiful?”
NVIDIA also revealed its next GPU—Blackwell Ultra—which is aimed at paving the way for the age of AI reasoning. It’s expected to be available in the second half of the year.
This also includes the introduction of the NVIDIA GB300 NVL72, which connects 72 Blackwell Ultra GPUs and 36 Arm Neoverse-based NVIDIA Grace CPUs in a rack-scale design, forming a single massive GPU. It outperforms the NVIDIA GB200 NVL72, delivering 1.5x more AI performance, and increasing revenue opportunity by 50x for AI factories.
“We designed Blackwell Ultra for this moment—it’s a single versatile platform that can easily and efficiently do pretraining, post-training, and reasoning AI inference,” said Huang.
The company also unveiled the NVIDIA DGX SuperPOD, which is built with NVIDIA Blackwell Ultra GPUs and makes the world’s most advanced enterprise AI infrastructure. It brings enterprises AI factory supercomputing with superior agentic AI reasoning. Huang said, “The NVIDIA Blackwell Ultra DGX SuperPOD provides out-of-the-box AI supercomputing for the age of agentic and physical AI.”
Paying tribute to astronomer Vera Rubin, Huang spoke about the upcoming next-generation NVIDIA Rubin Ultra GPU and NVIDIA Vera CPU architectures, which when paired together to form the Vera Rubin computing system, are expected to outperform the Blackwell architecture. Vera Rubin will come in the second half of 2026, with Vera Rubin Ultra coming in 2027.
Going one step further, Huang said the Vera Rubin system will be followed by the Feynman architecture (2028).
Describing Blackwell’s capabilities of supporting extreme scale-up he said, “The reason why we wanted to do this is to solve an extreme problem, and it’s called inference.”
NVIDIA Dynamo
To further accelerate inference on a large scale, the company is bringing NVIDIA Dynamo, an open-source modular inference framework to help accelerate and scale AI reasoning models in AI factories. It comes with the lowest cost and the highest efficiency. Huang said, “To enable a future of custom reasoning AI, NVIDIA Dynamo helps serve these models at scale, driving cost savings and efficiencies across AI factories.”
NVIDIA Blackwell RTX PRO
The company announced the NVIDIA RTX PRO Blackwell series, which consists of powerful workstation and server GPUs with boosted computing, AI inference, ray tracing, and neural rendering technologies, specially designed for AI, technical, creative, engineering, and design professionals. The lineup includes data center, desktop, and laptop GPUs.
Its features include NVIDIA streaming multiprocessor, fourth-generation RT cores, fifth-generation Tensor cores, larger and faster GDDR7 memory, Multi-Instance GPU (MIG), and more. The GPUs are expected to be available soon through various distribution partners, including Cisco, Dell, HP, Lenovo, and others.
Using NVIDIA NIM microservices—its optimized AI software—allow users to unlock greater possibilities and performance, allowing developers to build AI-powered app assistants, productivity tools, and content-creation workflows that integrate seamlessly.
NVIDIA Personal AI Computers
NVIDIA’s Project DIGITS is finally here—with a new name!
The company unveiled two personal AI supercomputers—DGX Spark and DGX Station—which are powered by the NVIDIA Blackwell Ultra platform. These machines, which aim to enable AI developers, researchers, data scientists, and students to prototype, fine-tune, and inference large models on desktops, will be developed by ASUS, Dell, HP Inc., and Lenovo.
DGX Spark is the world’s smallest AI supercomputer, while DGX Station brings data center-level performance to desktops for AI development.
The NVIDIA Quantum Computing Research Center
NVIDIA revealed it’s building a research center in Boston, Massachusetts, to provide cutting-edge technologies to advance quantum computing. It’s called the NVIDIA Accelerated Quantum Research Center, or NVAQC, and will integrate quantum hardware with AI supercomputers to solve quantum computing’s most challenging problems. “Quantum computing will augment AI supercomputers to tackle some of the world’s most important problems, from drug discovery to materials development,” said Jensen Huang.
NVIDIA’s Robotics Moves
Announcing Isaac GR00T N1, the world’s first open, customizable humanoid robot foundation model for generalized humanoid reasoning and skills, Jensen Huang proclaimed, “The age of generalist robotics is here!”
GR00T N1 forms the first of a family of fully customizable models that will be made available to robotics developers worldwide to help compensate for the shortage of skilled workers. “With NVIDIA Isaac GR00T N1 and new data-generation and robot-learning frameworks, robotics developers everywhere will open the next frontier in the age of AI.”
GR00T N1 comes with a dual-system architecture: “System 1” a fast-thinking action model to mirror human reflexes or intuition, and “System 2” a slow-thinking model for deliberate, methodical decision-making.
Adding to this is the NVIDIA Isaac GR00T Blueprint for generating synthetic data and Newton, an open-source physics engine for developing robots that is under development with Google DeepMind and Disney Research, which was exhibited on stage by Jensen Huang.
NVIDIA And General Motors
NVIDIA and GM will collaborate to build next-generation vehicles, factories and robots using AI, simulation and AI chips. This will include building custom AI systems using NVIDIA accelerated compute platforms, to optimize GM’s factory planning and robotics, as well as enhance in-vehicle hardware for future advanced driver-assistance systems (ADAS) for safe driving experiences.
This includes the NVIDIA Omniverse, a platform of APIs, SDKs, and services that enhance simulation workflows for industrial and robotic use cases. The Omniverse blueprint advances AI factory design and simulation.
“The era of physical AI is here, and together with GM, we’re transforming transportation, from vehicles to the factories where they’re made,” said Huang.
As such, NVIDIA is also releasing a massive, open-source dataset for building the next generation of physical AI, to help developers and researchers teach autonomous robots and vehicles how to interact with the physical world. Available on Hugging Face, the dataset offers developers 15 terabytes of data representing more than 320,000 trajectories for robotics training.
NVIDIA Halos
NVIDIA’s Halos operates at the intersection of autonomy and robotics, with a focus on the development of autonomous vehicles (AVs). It forms a comprehensive safety system combining automotive hardware and software safety solutions. From chips and software to tools and services, the holistic safety system works on three different levels.
“At the technology level, it spans platform, algorithmic, and ecosystem safety. At the development level, it includes design-time, deployment-time, and validation-time guardrails. And at the computational level, it spans AI training to deployment.” The system connects NVIDIA DGX for AI training, NVIDIA Omniverse and NVIDIA Cosmos for simulation, and NVIDIA DRIVE AGX for deployment.
It combines safety-focused technology research, development, deployment, partnerships, and collaborations by NVIDIA, including 15,000+ engineering years invested in vehicle safety, 10,000+ hours of contributions to international standards committees, 1,000+ AV-safety patents filed, and more.
“Halos’ holistic approach to safety is particularly critical in a setting where companies want to harness the power of generative AI for increasingly capable AV systems developed end to end, which preclude traditional compositional design and verification,” said Marco Pavone, lead AV researcher at NVIDIA.
NVIDIA, Alphabet, And Google
Building on their previous partnership, NVIDIA, Alphabet and Google revealed new initiatives to advance AI, democratize access to AI tools, speed up the development of physical AI, and transform industries including healthcare, manufacturing, and energy. This will include using AI and simulation technology to develop robots with grasping skills, reimagine drug discovery, optimize energy grids, and more.
“I’m really excited about the next phase of our partnership as we work together on agentic AI, robotics, and bringing the benefits of AI to more people around the world,” said Sundar Pichai, CEO of Google and Alphabet.
“It’s a great joy to see Google and NVIDIA researchers and engineers collaborate to solve incredible challenges, from drug discovery to robotics,” said Jensen Huang.
Google DeepMind and NVIDIA will work together to develop responsible AI and open models.
New Partnerships
Yum! Brands, the world’s largest restaurant chain and parent company of KFC, Taco Bell, Pizza Hut, and Habit Burger & Grill, is partnering with NVIDIA to deploy multiple AI solutions in over 500 restaurants this year. The company will also use NVIDIA computer vision software to enhance staffing solutions by analyzing drive-thru traffic and more.
Oracle and NVIDIA announced a first-of-its-kind integration between NVIDIA accelerated computing and inference software with Oracle’s AI infrastructure and generative AI services, to help organizations speed up the creation of agentic AI applications. This will be done by making 160+ AI tools and 100+ NVIDIA NIM microservices natively available through the OCI Console, among other moves.
What do you think about the reveals made so far at NVIDIA GTC 2025?
Let us know in the comments below!
First published on Wed, Mar 19, 2025
Liked what you read? That’s only the tip of the tech iceberg!
Explore our vast collection of tech articles including introductory guides, product reviews, trends and more, stay up to date with the latest news, relish thought-provoking interviews and the hottest AI blogs, and tickle your funny bone with hilarious tech memes!
Plus, get access to branded insights from industry-leading global brands through informative white papers, engaging case studies, in-depth reports, enlightening videos and exciting events and webinars.
Dive into TechDogs' treasure trove today and Know Your World of technology like never before!
Disclaimer - Reference to any specific product, software or entity does not constitute an endorsement or recommendation by TechDogs nor should any data or content published be relied upon. The views expressed by TechDogs' members and guests are their own and their appearance on our site does not imply an endorsement of them or any entity they represent. Views and opinions expressed by TechDogs' Authors are those of the Authors and do not necessarily reflect the view of TechDogs or any of its officials. While we aim to provide valuable and helpful information, some content on TechDogs' site may not have been thoroughly reviewed for every detail or aspect. We encourage users to verify any information independently where necessary.
Trending TD NewsDesk
Tech Titans’ AI Legal Storms: Microsoft, Google, OpenAI & X Mixed In Lawsuits
By TechDogs Bureau
Meta Reveals Video Editor, Threads Ads & Smart Glasses Updates Amid Criticism
By TechDogs Bureau
AI Under Scrutiny From Regulators Amid Missteps, Mistakes & Mounting Mistrust
By TechDogs Bureau
Marks & Spencer And Hertz Affected By The Latest Bout Of Cyber Incidents
By TechDogs Bureau
Fujitsu & RIKEN Unveil Superconducting Quantum Computer To Bridge Classical & Quantum Computing
By TechDogs Bureau
Join Our Newsletter
Get weekly news, engaging articles, and career tips-all free!
By subscribing to our newsletter, you're cool with our terms and conditions and agree to our Privacy Policy.
Join The Discussion